I recently had to move my entire cluster, and after bringing everything back online, I'm running into issues pulling images. For instance, when I tried to pull the homepage container, I got the following error: "Failed to pull image
2 Answers
You might want to try running `kubectl describe node ` as a troubleshooting step. My guess is that your CNI (Container Network Interface) might be having issues. Restarting the Cilium deployment in the kube-system namespace with `kubectl rollout restart deployment cilium -n kube-system` could also help.
It sounds like a network issue rather than a credential problem. Can you check if the host can reach the IP and port directly? That might give you some clues about what’s going on.
I thought so too, but other pods on the same node can pull images just fine. I'm using Talos Linux, so I can’t shell into the pods directly. I even tried deploying a debug pod with the ping command, but I couldn't pull the image. Any ideas on how I can check connectivity? I’m also seeing errors in the Talos dashboard about not being able to add a route, along with I/O timeouts for "discovery.talos.dev".